Amaury Lendasse

Contact

Information

P.O. Box 15400 Voice: +358 (0) 407 700237Department of Information and Computer Science Fax: +358 (0) 9 451 3277Aalto University School of Science E-mail: amaury.lendasse@aalto.fiFI-00076 Aalto FINLAND WWW: http://research.ics.aalto.fi/eiml

Date and Place

of Birth

April 16th, 1972, Tournai, Belgium

Citizenship and

Family Status

Belgian, married with Kati Pulkkinen, one daughter b. 2011

Research

Interests

Theory: Machine Learning, Time Series Prediction, Feature Selection, Functional Data AnalysisApplications: Chemometrics, Environmental Modeling, Corporate Finance, Internet Security

Education Université catholique de Louvain, Louvain-la-Neuve, Belgium

Ph.D. in Applied Sciences, October 2003• Dissertation Topic: “Analyse et prédiction de séries temporelles par méthodes non linéaires:

Application à des données industrielles et financières (Analysis and Prediction of Time Seriesusing Nonlinear Methods: Application to Industrial and Financial Datasets), Presses univer-sitaires de Louvain, ISBN: 2930344342, Louvain-la-Neuve (Belgium)”

• Advisors: Michel Verleysen and Vincent WertzMaster’s degree in Control (Diplôme d’étude complémentaire en automatique), June 1997with Magna Cum Laude

Master’s degree in Mechanics (Ingénieur Civil en mécanique), June 1996 with Cum Laude

Languages 1. French, 2. English (fluent), 3. Finnish (basic)

Employment University of Iowa, Iowa, USAAssociate Professor May 2014 - present

Arcada University of Applied Sciences, Helsinki, FinlandVisiting Position January 2014 -present

Ikerbasque, San Sebastian, SpainResearch Professor (tenured position) June 2013 -July 2014

Aalto University (former HUT), Otaniemi, FinlandAdjunct Professor, Docent and Chief Research Scientist Apr 2007 - presentIncludes supervision of Ph.D. Students and Master Students.Group Leader Jan 2005 - Mar 2007Creation of the "Environmental and Industrial Machine Learning" research groupPostdoctoral Researcher Jan - Dec 2004Working on a research project in collaboration with Nokia

University of Memphis, Tennessee, USAPostdoctoral Researcher Oct - Dec, 2003Postdoctoral researcher at the Computational Neurodynamics laboratoryin collaboration with NASA

Université catholique de Louvain, Louvain-la-Neuve, BelgiumPh.D. Student Oct 1998 - Sep 2003Course Assistant (university staff) and Ph.D. studentin the Automatic Control LaboratoryPh.D. Student Sep 1997 - Sep 1998Scholarship from Université catholique de Louvain (FSR)in the Electronics Laboratory

Invited Positions Invited Professor at the Ecole Centrale de Lille (one month in 2011, 2012 and 2013)

Invited Professor at the Universite de Lille (one month in 2012)

Invited Professor at the Universite Paris I, Sorbonne (one month in 2011 and in 2012)

Invited Professor at the University of the Basque Country (one month in 2012)

Supervised Ph.D.

Theses

Qi Yu, Machine Learning for Bankruptcy prediction Problems, 2013, Aalto University School ofScience and Technology, Finland

Yoan Miché, Developing Fast Machine Learning Techniques with Applications to Steganalysis Prob-lems, double degree Ph.D. thesis, co-supervised by Christian Jutten from the INPG-France, 2010,Aalto University School of Science and Technology, Finland

Elia Liitiäinen, Noise Variance Estimation for Function Approximation, 2010, Aalto UniversitySchool of Science and Technology, Finland

Antti Sorjamaa, The Problem of Missing Data in Spatio-Temporal Databases, 2010, Aalto UniversitySchool of Science and Technology, Finland

Francesco Corona, Development and Application of Data-Derived Models for Monitoring IndustrialProcesses, 2006, Università Degli Studi Di Cagliari, Italy

Supervising Ph.D.

Theses

Emil Eirola, Using causal relationships for ensemble modelling of the Baltic Sea (started in 2009)

Dusan Sovilj, Machine Learning Methods for Environmental Modeling of Baltic Sea (started in 2009)

Alexander Grigorievskiy, Machine Learning for Environmental Modeling (Prediction of IncompleteEnvironmental Time Series) (started in 2013)

Member of Ph.D.

Juries

Francis wyffels, Sequence Generation with Reservoir Computing Systems, 2013, University of Gent,Belgium

Gines Rubio Flores, Advanced Computational Intelligence Models for functional approximation andprediction of time series in parallel architectures, 2010, Universidad de Granada, Spain

Federico Montesino Pouzols, Mining and Control of Network Traffic by Computational Intelligence,2009, University of Sevilla, Spain

Rui Nian, Pattern Recognition with Statistical and Geometrical Mutual Relationships, 2008, OceanUniversity of China (Qingdao), China

Alberto Guillen, Design of intelligent systems in parallel computation platforms, 2007, Universidadde Granada, Spain

Luis Javier Herrera, Intelligent and Adaptive system for function approximation and time seriesprediction using advances models, 2007, Universidad de Granada, Spain.

Damien Francois, High-dimensional data analysis: optimal metrics and feature selection, 2007, Uni-versité catholique de Louvain, Belgium

Amaury Lendasse: Summary of Merits

Teaching Machine Learning: Basic Principles (T-61.3050), 2012Information visualization (T-61.5010), 2010-2012Bankruptcy Prediction at the Ecole Centrale de Lille, 2010-Statistical Signal Modeling (T-61.3040), 2010-2012Machine Learning for Corporate Finance (T-61.9910), 2011Bankruptcy Prediction (T-61.6020), 2009Time Series Analysis and Modeling of Environmental Data, 2009 (BONUS+ EEIG course)High-dimensional Data Analysis: From Optimal Metrics to Feature Selection (T-61.6010), 2008Nonlinear Dimensionality Reduction (T-61.6050), 2007Introductory Elements of Functional Data Analysis (T-61.6030), 2007Variable Selection for Regression (T-61.6040), 2006Neural Networks for Modeling and Control of Dynamic Systems (T-61.6050), 2005Regularization and Sparse Basis Function Approximations (T-122.102), 2005Support Vector Machines (T-61.190), 2005Analysis of Time Series and Sequences (T-122.101), 2004

Supervised Ph.DTheses

Instructor for 9 Ph.D Theses at the Aalto University School of Science.

Supervised M.ScTheses

Instructor for 13 M.Sc Theses at the Aalto University School of Science. Instructor for 15 M.ScTheses at the Universite catholique de Louvain in Belgium under the supervision of Prof. Verleysen

Seminar andInvited Talks

Jun.’13, Keynote Speaker at IWANN’13, "Extreme Learning Machine: A Robust Modeling Tech-nique? Yes!", Tenerife (SP)Apr.’12, Keynote Speaker at Statlearn’12, "Challenging problems in Statistical Learning", Lille (FR)Sept.’11, University of the Basque Country, Dept. of Computer Science, invited by Prof. GranaJul.’10, Univ. de Granada, E.T.S. Ingenierías Informática y Telecom, invited by Prof. PrietoJun.’08, Institute for Robotics and Cognitive Systems, Univ. Luebeck, invited by Prof. SchweikardJun.’07, Univ. de Granada, E.T.S. Ingenierías Informática y Telecom, invited by Prof. RojasMay’05, Univ. Paris 1 - Panthéon-Sorbonne (France), Lab. SAMOS, invited by Prof. CottrellMay’04, Invited Lecturer at the University of Tartu (Estonia), Computer Sciences DepartmentApr.’03, Univ. of Memphis (USA), Comp. Neurodynamics Laboratory, invited by Prof. KozmaMay’02, Ohio State Univ. (USA), Collaborative Center of Control Science, invited by Prof. YurkovichMay’02, Ohio Univ. (USA), School of Electrical Eng. & Comp. Science, invited by Prof. J. ZhuMar.’00, Katholieke Univ. Leuven (Belgium), ESAT Laboratory, invited by Prof. Suykens

Other ScientificActivity

Member of the EC’s Network of Excellence under Framework 6 PASCAL and Framework 7 PASCAL2(Pattern Analysis, Statistical modeling and ComputAtional Learning)Member of the board of the Helsinki Graduate School in Computer Science and EngineeringIEEE member, Vice-Chair for Europe of the Standards Committee of the IEEE ComputationalIntelligence Society (CIS) (2011-2012)

Reviewer forScientificJournals

Neurocomputing (also Guest Editor), IEEE Trans. on Neural Networks, Neural Processing Letters,Journal of Chemometrics, Computational Statistics and Data Analysis, Information Sciences, Inter-national Journal of Forecasting (also Guest Editor), International Journal of Pattern Recognitionand Artificial Intelligence, Case Studies in Business, Industry and Government Statistics (also As-sociate Editor), Systems, Man and Cybernetics - Part B, Fuzzy Sets and Systems, Neural Networks,Evolving Systems, Transactions on Knowledge and Data Engineering, BMC Biotechnology

ConferenceCommittees

International Symposium on Extreme Learning Machine (ELM’11 to ’14) (also Program Chair)International Conference on Artificial Neural Networks (ICANN’11) (also Organizing Chair)International Symposium on Extreme Learning Machines (ELM’11) (also Publicity Chair)European Symposium on Neural Networks, Computational Intelligence and Machine Learning (ESANN’11to ’14)International Joint Conference on Neural Networks (WCCI’10/ IJCNN’10) (also Workshops Chair)Workshop on Self-Organizing Maps (WSOM’05 and ’09) (also Special Session Chair)International and Interdisciplinary Conf. on Adapt. Knowledge Representation & Reasoning (AKRR’08)International Work-Conference on Artificial Neural Networks (IWANN’07)Computational Methods for Modelling and learning in Social and Human Sciences (MASHS’07-’10)International Conference on Natural Computing (NC’07)International Joint Conference on Neural Networks (IJCNN’04)

OrganizedEvents

Organizer of the the special session: "Machine learning techniques based on random projections" atthe European Symposium on Neural Networks (ESANN’10)Organizer of the European Symposium on Time Series Prediction (ESTSP’07 and ESTSP’08)Editor of the Proceedings of European Symposium on Time Series Prediction (ESTSP’07 and ESTSP’08)Organizer of the special session: ”Time Series Competition: The CATS Benchmark” at the Interna-tional Joined Conference on Neural Networks 2004 (IJCNN’04)Organizer of the special session: ”Feature Selection and Dimension Reduction for Regression” at theInternational Conference on Artificial Neural Networks 2006 (ICANN’06)

Award Best Paper Award at the NN3 Neural Network Forecasting Competition at the International JoinedConference on Neural Networks 2007 (IJCNN’07).

ScientificPublications

51 articles published in international journals, 130 refereed conference papers and book chapters, 3Books (including PhD Thesis, in French), See attached complete list of publications. H-index: 24(24 published papers with at least 24 citations each) using Google Scholar, h-index: 15 using Scopus,Erdös number: 3

Industrial andEuropeanProjects

Leading Scientist in the "Tivit Data 2 Intelligence Program" (D2I) funded by TEKES, in collabo-ration with F-Secure, funding: 214 KEUR in 2014, 100 KEUR in 2013 and 120 KEUR in 2012.Leading Scientist in the "Tivit Future Internet Program" (FI-SHOK) funded by TEKES, in collab-oration with F-Secure, funding 240 KEUR in 2011.Supervisor of Dr. Federico Pouzols, Marie Curie Fellow of the Intra-European Fellowship (IEF)programme, within the EU’s Seventh Framework Programme FP7, funding 170 KEUR, 2009-2011.Leading scientist for the ”Assessment and Modelling of Baltic Ecosystem Response (AMBER)”project, within the EU’s Seventh Framework Programme FP7, funding: 50 KEUR in 2010.Leading scientist for the ”Nonlinear temporal and spatial forecasting: modeling and uncertainty anal-ysis, Phase II (NOTES-2)” project, funded by TEKES, Modeling and Simulation (MASI) Program,2008. Tekes’ funding share: 401 KEUR.Leading scientist for the ”Nonlinear temporal and spatial forecasting: modeling and uncertaintyanalysis (NOTES)” project, funded by TEKES, Modeling and Simulation (MASI) Program, 2006-2007. Tekes’ funding share: 602 KEUR.Head Scientist and chairman of the steering Group for the ”Developing Chemometrics with the Toolsof Information Sciences (CHESS)” project, funded by TEKES, Modeling and Simulation (MASI)Program, 2006-2007. Tekes’ funding share: 315 KEUR.

Grad. SchoolGrantsco-obtained withStudents

3 PhD students and 2 former PhD students funded by the Helsinki Graduate School in ComputerScience and Engineering (HECSE). Total fundings: 375 KEUR1 PhD student funded by the Finnish Doctoral Programme in Computational Sciences (FICS). Totalfundings: 75 KEUR

A. Lendasse, Bibliometrics Summary and LAST Publications

All the publications are publicly available at http://users.ics.aalto.fi/lendasse/.

Bibliometric data:

• Number of published peer-reviewed journal articles: 51

• Number of published peer-reviewed conference articles: 130

• Number of published books (monographs, edited collections): 2

Bibliometric indicators:

• Total number of publications: 230 (Google Scholar), 114 (Scopus)

• Total number of citations: 2510 (Google Scholar), 887 (Scopus)

• h-index: 24 (Google Scholar), 15 (Scopus)

NEW publications:

1. Long-term Time Series Prediction using OP-ELM. Alexander Grigorievskiy, Yoan Miche, Anne-Mari Ventela and Amaury Lendasse. In Neural Networks, 2013, to appear.(5-Year Impact Factor: 2.501)

2. ELMVIS: a Nonlinear Visualization Technique using Random Permutations and Extreme LearningMachine, Anton Akusok, Amaury Lendasse and Yoan Miche. Accepted in IEEE Intelligent Systems.(5-Year Impact Factor: 2.538)

The evolution of the number of citations according to Scopus is given in Figure 1.

Figure 1: Number of Citations from 2004 to 2013 (according to Scopus)

1

Ranking (based on my personal opinion and not according to the

number of citations)

[1] Y. Miche, A. Sorjamaa, P. Bas, O. Simula, C. Jutten, and A. Lendasse, “OP-ELM: Optimally-pruned extreme learning machine,” IEEE Transactions on Neural Networks, vol. 21, pp. 158–162,January 2010.This paper is presenting an improvement of the Extreme Learning Machine. Extreme LearningMachine is the fastest way to build nonlinear regressors and classifiers. OP-ELM is an elegantway to make ELM robust and efficient. In this paper, we have shown that OP-ELM is providingthe best ratio efficiency / computational time among all the state-of-the-art methods in MachineLearning (SVM, GP, MLP, ELM, etc.) This method is now used in many fields like Steganalysis,Environmental Modeling, Time Series Prediction, Bankruptcy Prediction, etc.Number of citations: 72/130/20 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 3.37DOI: http://dx.doi.org/10.1109/TNN.2009.2036259.

[2] A. Lendasse, J. A. Lee, V. Wertz, and M. Verleysen, “Forecasting electricity consumption usingnonlinear projection and self-organizing maps,” Neurocomputing, vol. 48, pp. 299–311, October2002.In this paper, a new methodology to select and/or build the inputs of a prediction model is presented.This methodology is using curvilinear component analysis in order to select the necessary andsufficient information to predict the future of a Time Series. Kohonen Maps are used as predictionmodels due to their robustness. This methodology is now recognized as one of the most efficient forthe prediction of electricity consumption.Number of citations: 35/51/23 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 1.60URL: http://www.sciencedirect.com/science/article/pii/S0925231201006464.

[3] J. A. Lee, A. Lendasse, and M. Verleysen, “Nonlinear projection with curvilinear distances: Isomapversus curvilinear distance analysis,” Neurocomputing, vol. 57, pp. 49–76, March 2004.Together with my former master student John Lee, we have developed a new method for nonlinearprojection: Curvilinear Distance Analysis (CDA). Like ISOMAP, CDA is using geodesic distances,but we have shown that CDA is more robust that ISOMAP. CDA is now one of the most usedmethod for dimensionally reduction.Number of citations: 82/131/61 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 1.60DOI: http://dx.doi.org/doi:10.1016/j.neucom.2004.01.007.

[4] F. Rossi, A. Lendasse, D. Francois, V. Wertz, and M. Verleysen, “Mutual information for the selectionof relevant variables in spectrometric nonlinear modelling,” Chemometrics and Intelligent LaboratorySystems, vol. 80, pp. 215–226, February 2006.In this paper, we have developed a methodology to select input variable in the field of spectrometricmodeling. This approach is new in the field of Chemometrics and give the possibility to improvethe interpretability of such spectrometric problems. The methodology is using efficiently MutualInformation in order to select the most relevant variables among several thousand of candidates.This paper is now one of the reference papers in Chemometrics.Number of citations: 94/142/43 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 2.30DOI: http://dx.doi.org/doi:10.1016/j.chemolab.2005.06.010.

[5] A. Sorjamaa, J. Hao, N. Reyhani, Y. Ji, and A. Lendasse, “Methodology for long-term prediction oftime series,” Neurocomputing, vol. 70, pp. 2861–2869, October 2007.This paper is joint work of all my former master students and myself. We have compared thetwo main methodologies for the longterm prediction of Time Series: the recursive and the directmethodologies. This paper is important since we have shown that the direct approach is the mostaccurate and the most robust. Since the publication of this paper, the number of works using therecursive methodology has decreased considerably.

2

Number of citations: 67/114/40 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 1.60DOI: http://dx.doi.org/10.1016/j.neucom.2006.06.015.

[6] G. Simon, A. Lendasse, M. Cottrell, J.-C. Fort, and M. Verleysen, “Time series forecasting: Obtain-ing long term trends with self-organizing maps,” Pattern Recognition Letters, vol. 26, pp. 1795–1808,September 2005.This paper is a joint work with my former student G. Simon. We have presented a new method topredict the very long trend of a Time Series using a double Self-Organizing Map. This methodologyis useful to predict economical Time Series like the price of natural gas.Number of citations: 26/34/14 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 1.72DOI: http://dx.doi.org/doi:10.1016/j.patrec.2005.03.002.

[7] A. Lendasse, D. Francois, V. Wertz, and M. Verleysen, “Vector quantization: a weighted version fortime-series forecasting,” Future Generation Computer Systems, vol. 21, no. 7, pp. 1056–1067, 2005.In this paper, I have presented a new method to predict Time Series using Vector Quantization Theadvantage of the method is that it is possible to determine automatically the relative importanceof each input variable used for the prediction. The method is now used by other authors in orderto improve the performance of other prediction methods using this relative importance as a prepro-cessing.Number of citations: 11/16/9 (according to Scopus/Google-Scholar/Microsoft-Academic)Impact Factor: 1.98DOI: http://dx.doi.org/10.1016/j.future.2004.03.006.

[8] A. Lendasse, G. Simon, V. Wertz, and M. Verleysen, “Fast bootstrap methodology for regressionmodel selection,” Neurocomputing, vol. 64, pp. 161–181, March 2005.This paper is presenting an improvement of the well-know bootstrap method for model structureselection. The bootstrap is very efficient to estimate the performance of a model but is very slowsince thousands of bootstrap repetitions are usually needed. Using some properties of the optimismcalculated by the bootstrapping, we have divided the computational time by several orders of magni-tude. This methodology has been illustrated with LS-SVM and RBFN. Other authors have later-onapplied the Fast-Bootstrap to other nonlinear models.Number of citations: 8/17/17 (according to Scopus/Google-Scholar/Microsoft-Academic)5-Year Impact Factor: 1.60DOI: http://dx.doi.org/doi:10.1016/j.neucom.2004.11.017.

[9] A.-M. Ventelä, T. Kirkkala, A. Lendasse, M. Tarvainen, H. Helminen, and J. Sarvala, “Climate-related challenges in long-term management of säkylän pyhäjärvi (SW finland),” Hydrobiologia,vol. 660, pp. 49–58, 2011.This paper is a new publication in the field of Environmental Modeling. This an invited paperthat shows how to use new and efficient variable selection methods in order to understand complexenvironmental systems. Number of citations: 5/3/0 (according to Scopus/Google-Scholar/Microsoft-Academic) 5-Year Impact Factor: 2.07DOI: http://dx.doi.org/10.1007/s10750-010-0415-4.

[10] A. Lendasse, E. de Bodt, V. Wertz, and M. Verleysen, “Nonlinear financial time series forecasting -application to the bel 20 stock market index,” European Journal of Economic and Social Systems,vol. 14, pp. 81–92, February 2001.This paper is my first journal paper. It presents a new method to predict Financial Time Series.I have shown that it is possible to predict accurately the sign of the return of a financial TimeSeries (with 75% of correct prediction of the sign). We can then answer to the following question:is the value of a stock market going to increase or decrease during the next few days? The resultsalso show that stock markets related to small countries are easer to predict since the hypothesis offinancial efficiency is not totally valid in these countries.Number of citations: 81/35 (according to Google-Scholar/Microsoft-Academic)WEB: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.11.9681.

3

Amaury Lendasse, FULL Publication List on June 2014

Journal articles

[1] Emil Eirola, Gauthier Doquire, Michel Verleysen and Amaury Lendasse. Distance estimation innumerical data sets with missing values. Information Sciences, volume 240:pages 115–128, 2013.

[2] Benoît Frénay, Mark van Heeswijk, Yoan Miche, Michel Verleysen and Amaury Lendasse. Featureselection for nonlinear models using extreme learning machines. Neurocomputing, volume 102:pages111–124, 2013.

[3] Alexander Grigorievskiy, Yoan Miche, Anne-Mari Ventelä and Amaury Lendasse. Long-term timeseries prediction using OP-ELM. Cognitive Computation, 2013. To appear.

[4] Bo He, Dongxun Xu, Rui Nian, Mark van Heeswijk, Qi Yu, Yoan Miche and Amaury Lendasse. Fastface recognition via sparse coding and extreme learning machine. Cognitive Computation, 2013. Toappear.

[5] Yoan Miche, Anton Akusok, József Hegedüs, Rui Nian and Amaury Lendasse. A two-stage method-ology using k-NN and false positive minimizing ELM for nominal data classification. CognitiveComputation, 2013. To appear.

[6] Yoan Miche, Meng-Hiot Lim, Amaury Lendasse and Yew-Soon Ong. Meme representations for gameagents. World Wide Web, pages 1–20, 2013.

[7] Rui Nian, Bo He, Bing Zheng, Mark van Heeswijk, Qi Yu, Yoan Miche and Amaury Lendasse.Extreme learning machine towards dynamic model hypothesis in fish ethology research. Neurocom-puting, 2013. To appear.

[8] Mark van Heeswijk, Qi Yu, Rui Nian, Bo He, Yoan Miche and Amaury Lendasse. BIP(CV)-ELM:Effective and adaptive pretraining method for extreme learning machines. Cognitive Computation,2013. To appear.

[9] Qi Yu, Yoan Miche, Emil Eirola, Mark van Heeswijk, Eric Séverin and Amaury Lendasse. Regularizedextreme learning machine for regression with missing data. Neurocomputing, volume 102:pages 45–51, 2013.

[10] Qi Yu, Yoan Miche, Eric Séverin and Amaury Lendasse. Bankruptcy prediction using extremelearning machine and financial expertise. Neurocomputing, 2013. To appear.

[11] Qi Yu, Mark van Heeswijk, Yoan Miche, Rui Nian, Bo He, Eric Séverin and Amaury Lendasse.Ensemble delta test- extreme learning machine (DT-ELM) for regression. Cognitive Computation,2013. To appear.

[12] Rui Nian, Bo He and Amaury Lendasse. 3d object recognition based on a geometrical topology modeland extreme learning machine. Neural Computing and Applications, pages 1–7, 2012, To appear.ISSN 0941-0643.

[13] Federico Montesino Pouzols and Amaury Lendasse. Adaptive kernel smoothing regression for spatio-temporal environmental datasets. Neurocomputing, volume 90:pages 59–65, August 2012. ISSN0925-2312.

[14] Laura Kainulainen, Yoan Miche, Emil Eirola, Qi Yu, Benoit Frénay, Eric Séverin and AmauryLendasse. Ensembles of local linear models for bankruptcy analysis and prediction. Case Studies inBusiness, Industry and Government Statistics (CSBIGS), volume 4(2), November 2011.

[15] E. Liitiäinen, F. Corona and A. Lendasse. On the curse of dimensionality in supervised learning ofsmooth regression functions. Neural Processing Letters, volume 34(2):pages 133–154, 2011.

1

[16] Yoan Miche, Mark van Heeswijk, Patrick Bas, Olli Simula and Amaury Lendasse. TROP-ELM: a double-regularized ELM using LARS and tikhonov regularization. Neurocomputing, vol-ume 74(16):pages 2413–2421, September 2011.

[17] Mark van Heeswijk, Yoan Miche, Erkki Oja and Amaury Lendasse. GPU-accelerated and paral-lelized ELM ensembles for large-scale regression. Neurocomputing, volume 74(16):pages 2430–2437,September 2011.

[18] Francesco Corona, Amaury Lendasse and Elia Liitiäinen. A boundary corrected expansion of themoments of nearest neighbor distributions. Random Structures and Algorithms, volume 37(2):pages223–247, September 2010.

[19] Alberto Guillén, Luis Herrera, Gines Rubio, Amaury Lendasse and Hector Pomares. New method forinstance or prototype selection using mutual information in time series prediction. Neurocomputing,volume 73(10–12):pages 2030–2038, June 2010.

[20] Amaury Lendasse, Timo Honkela and Olli Simula. European symposium on times series prediction.Neurocomputing, volume 73(10–12):pages 1919–1922, June 2010.

[21] Elia Liitiäinen, Amaury Lendasse and Francesco Corona. Residual variance estimation using anearest neighbor statistic. Journal of Multivariate Analysis, volume 101(4):pages 811–823, April2010.

[22] Paul Merlin, Antti Sorjamaa, Bertrand Maillet and Amaury Lendasse. X-SOM and l-SOM: Adouble classification approach for missing value imputation. Neurocomputing, volume 73(7-9):pages1103–1108, March 2010.

[23] Yoan Miche, Antti Sorjamaa, Patrick Bas, Olli Simula, Christian Jutten and Amaury Lendasse.OP-ELM: Optimally-pruned extreme learning machine. IEEE Transactions on Neural Networks,volume 21(1):pages 158–162, January 2010.

[24] Federico Montesino Pouzols and Amaury Lendasse. Evolving fuzzy optimally pruned extreme learningmachine for regression problems. Evolving Systems, volume 1(1):pages 43–58, August 2010.

[25] Federico Montesino Pouzols, Amaury Lendasse and Angel Barriga Barros. Autoregressive time seriesprediction by means of fuzzy inference systems using nonparametric residual variance estimation.Fuzzy Sets and Systems, volume 161(4):pages 471–497, February 2010.

[26] Antti Sorjamaa, Amaury Lendasse, Yves Cornet and Eric Deleersnijder. An improved method-ology for filling missing values in spatiotemporal climate data set. Computational Geosciences,volume 14:pages 55–64, January 2010.

[27] Anne-Mari Ventelä, Teija Kirkkala, Amaury Lendasse, Marjo Tarvainen, Harri Helminen and JoukoSarvala. Climate-related challenges in long-term management of säkylän pyhäjärvi (SW finland).Hydrobiologia, volume Online First:pages 1–10, 2010.

[28] Qi Yu, Yoan Miche, Antti Sorjamaa, Alberto Guillén, Amaury Lendasse and Eric Séverin. OP-KNN:Method and applications. Advances in Artificial Neural Systems, volume 2010(597373):page 6 pages,February 2010.

[29] Francesco Corona, Elia Liitiäinen, Amaury Lendasse, Lorenzo Sassu, Stefano Melis and RobertoBaratti. A SOM-based approach to estimating product properties from spectroscopic measurements.Neurocomputing, volume 73(1–3):pages 71–79, December 2009.

[30] Elia Liitiäinen, Michel Verleysen, Francesco Corona and Amaury Lendasse. Residual variance esti-mation in machine learning. Neurocomputing, volume 72(16–18):pages 3692–3703, October 2009.

[31] Yoan Miche, Patrick Bas, Amaury Lendasse, Christian Jutten and Olli Simula. A feature selectionmethodology for steganalysis. Traitement du Signal, volume 26(1):pages 13–30, May 2009.

2

[32] Yoan Miche, Patrick Bas, Amaury Lendasse, Christian Jutten and Olli Simula. Reliable steganalysisusing a minimum set of samples and features. EURASIP Journal on Information Security, volume2009(1):pages 1–13 (Article ID 901381), March 2009.

[33] Antti Sorjamaa, Paul Merlin, Bertrand Maillet and Amaury Lendasse. A non-linear approach forcompleting missing values in temporal databases. European Journal of Economic and Social Systems,volume 22(1):pages 99–117, November 2009.

[34] Francesco Corona, Satu-Pia Reinikainen, Kari Aaljoki, Annikki Perkkiö, Elia Liitiäinen, RobertoBaratti, Amaury Lendasse and Olli Simula. Wavelength selection using the measure of topologicalrelevance on the self-organizing map. Journal of Chemometrics, volume 22(11–12):pages 610–620,November-December 2008.

[35] Alberto Guillén, DuŽan Sovilj, Fernando Mateo, Ignacio Rojas and Amaury Lendasse. Minimizingthe delta test for variable selection in regression problems. International Journal of High PerformanceSystems Architecture, volume 1(4):pages 269–281, 2008.

[36] Tuomas Kärnä, Francesco Corona and Amaury Lendasse. Gaussian basis functions for chemomet-rics. Journal of Chemometrics, volume 22(11–12):pages 701–707, November-December 2008.

[37] Elia Liitiäinen, Francesco Corona and Amaury Lendasse. On non-parametric residual varianceestimation. Neural Processing Letters, volume 28(3):pages 155–167, December 2008.

[38] Elia Liitiäinen, Amaury Lendasse and Francesco Corona. Bounds on the mean power-weightednearest neighbour distance. Proceedings of the Royal Society A, volume 464(2097):pages 2293–2301,September 2008.

[39] Amaury Lendasse, Erkki Oja, Olli Simula and Michel Verleysen. Time series prediction competition:The CATS benchmark. Neurocomputing, volume 70(13-15):pages 2325–2329, August 2007.

[40] Antti Sorjamaa, Jin Hao, Nima Reyhani, Yongnan Ji and Amaury Lendasse. Methodology for long-term prediction of time series. Neurocomputing, volume 70(16-18):pages 2861–2869, October 2007.

[41] Fabrice Rossi, Amaury Lendasse, Damien François, Vincent Wertz and Michel Verleysen. Mutual in-formation for the selection of relevant variables in spectrometric nonlinear modelling. Chemometricsand Intelligent Laboratory Systems, volume 80(2):pages 215–226, February 2006.

[42] Amaury Lendasse, Damien Francois, Vincent Wertz and Michel Verleysen. Vector quantiza-tion: a weighted version for time-series forecasting. Future Generation Computer Systems, vol-ume 21(7):pages 1056–1067, 2005. ISSN 0167-739X.

[43] Amaury Lendasse, Geoffroy Simon, Vincent Wertz and Michel Verleysen. Fast bootstrap methodologyfor regression model selection. Neurocomputing, volume 64:pages 161–181, March 2005.

[44] Geoffroy Simon, Amaury Lendasse, Marie Cottrell, Jean-Claude Fort and Michel Verleysen. Timeseries forecasting: Obtaining long term trends with self-organizing maps. Pattern Recognition Letters,volume 26(12):pages 1795–1808, September 2005.

[45] Eric de Bodt, Amaury Lendasse, Pierre Cardon and Michel Verleysen. Self-organizing feature mapsfor the classification of investment funds. Journal of Economic and Social Systems, volume 17(1-2):pages 183–195, 2004.

[46] John A. Lee, Amaury Lendasse and Michel Verleysen. Nonlinear projection with curvilinear dis-tances: Isomap versus curvilinear distance analysis. Neurocomputing, volume 57:pages 49–76, March2004.

[47] Geoffroy Simon, Amaury Lendasse, Marie Cottrell, Jean-Claude Fort and Michel Verleysen. Doublequantization of the regressor space for long-term time series prediction: Method and proof of stability.Neural Networks, volume 17(8-9):pages 1169–1181, October-November 2004. Special Issue.

3

[48] Amaury Lendasse, John A. Lee, Vincent Wertz and Michel Verleysen. Forecasting electricityconsumption using nonlinear projection and self-organizing maps. Neurocomputing, volume 48(1-4):pages 299–311, October 2002.

[49] Amaury Lendasse, Eric de Bodt, Vincent Wertz and Michel Verleysen. Nonlinear financial timeseries forecasting - application to the bel 20 stock market index. European Journal of Economic andSocial Systems, volume 14(1):pages 81–92, February 2001.

[50] Amaury Lendasse, John A. Lee, Vincent Wertz, Eric de Bodt and Michel Verleysen. Dimensionreduction of technical indicators for the prediction of financial time series, application to the bel 20market index. European Journal of Economic and Social Systems, volume 15(2):pages 31–48, 2001.

Conference papers

[51] Amaury Lendasse, Anton Akusok, Olli Simula, Francesco Corona, Mark van Heeswijk, Emil Eirolaand Yoan Miche. Extreme learning machine: A robust modeling technique? yes! In Joan CabestanyIgnacio Rojas, Gonzalo Joya, editor, IWANN 2013, Part I, LNCS, volume 7902, pages 17–36.Springer Heidelberg, Tenerife, Spain, June 12-14 2013. To appear.

[52] Dusan Sovilj, Amaury Lendasse and Olli Simula. Extending extreme learning machine with combi-nation layer. In Joan Cabestany Ignacio Rojas, Gonzalo Joya, editor, IWANN 2013, Part I, LNCS,volume 7902, pages 417–426. Springer Heidelberg, Tenerife, Spain, June 12-14 2013. To appear.

[53] Emil Eirola, Amaury Lendasse, Vincent Vandewalle and Christophe Biernacki. Mixture of gaussiansfor distance estimation with missing data. In Machine Learning Reports 03/2012, pages 37–45.2012. Proceedings of the Workshop - New Challenges in Neural Computation 2012.

[54] Andrej Gisbrecht, Dusan Sovilj, Barbara Hammer and Amaury Lendasse. Relevance learning fortime series inspection. In ESANN’12, pages 489–494. April 2012.

[55] Alberto Guillén, Dusan Sovilj, Mark van Heeswijk, Luis Javier Herrera, Amaury Lendasse, HectorPomares and Ignacio Rojas. Evolutive Approaches for Variable Selection Using a Non-parametricNoise Estimator, Studies in Computational Intelligence, volume 415, pages 243–266. SpringerBerlin Heidelberg, 2012.

[56] József Hegedüs, Yoan Miche, Alexander Ilin and Amaury Lendasse. Methodology for behavioral-based malware analysis and detection using random projections and k-nearest neighbors classifiers.In 7th International Conference on Computational Intelligence and Security (CIS2011). Sanya,China, December 2011.

[57] József Hegedüs, Yoan Miche, Alexander Ilin and Amaury Lendasse. Random projection method forscalable malware classification. In 14th International Symposium on Recent Advances in IntrusionDetection. California, USA, September 2011.

[58] Federico Montesino Pouzols and Amaury Lendasse. Adaptive kernel smoothing regression for spatio-temporal environmental datasets. In Michel Verleysen, editor, ESANN2011: 19th European Sym-posium on Artificial Neural Networks, Computational Intelligence and Machine Learning, pages87–92. d-side Publications, Bruges, Belgium, April 27–29 2011.

[59] Li Yao, Amaury Lendasse and Francesco Corona. Locating anomalies using bayesian factorizationsand masks. In Michel Verleysen, editor, ESANN2011: 19th European Symposium on ArtificialNeural Networks, Computational Intelligence and Machine Learning, pages 207–212. d-side Publi-cations, Bruges, Belgium, April 27–29 2011.

[60] Zhanxing Zhu, Francesco Corona, Amaury Lendasse, Roberto Baratti and Jose Romagnoli. Locallinear regression for soft-sensor design with application to an industrial deethanizer. In 18th WorldCongress of the International Federation of Automatic Control (IFAC). Milano, Italy, August 2011.

4

[61] Zhanxing Zhu, Francesco Corona, Amaury Lendasse, Roberto Baratti and Jose Romagnoli. Locallinear regression for soft-sensor design with application to an industrial deethanizer. In IFAC2011Eighteenth World Congress of the International Federation of Automatic Control, pages 2839–2844.Milan, Italy, August 28–September 2 2011.

[62] Francesco Corona, Elia Liitiäinen, Amaury Lendasse, Roberto Baratti and Lorenzo Sassu. A con-tinuous regression function for the delaunay calibration method. In Proceedings of IFAC/DYCOPS2010 9th International Symposium on Dynamics and Control of Process Systems, Leuven (Bel-gium), pages 180–185. July 5-7 2010.

[63] Laura Kainulainen, Qi Yu, Yoan Miche, Emil Eirola, Eric Séverin and Amaury Lendasse. Ensemblesof locally linear models: Application to bankruptcy prediction. In Robert Stahlbock and Sven F.Crone, editors, Proceedings of the 2010 International Conference on Data Mining, pages 280–286.Worldcomp1́0, July 2010.

[64] Yoan Miche, Emil Eirola, Patrick Bas, Olli Simula, Christian Jutten, Amaury Lendasse and MichelVerleysen. Ensemble modeling with a constrained linear system of leave-one-out outputs. In MichelVerleysen, editor, ESANN2010: 18th European Symposium on Artificial Neural Networks, Com-putational Intelligence and Machine Learning, pages 19–24. d-side Publications, Bruges, Belgium,April 28–30 2010.

[65] Yoan Miche, Benjamin Schrauwen and Amaury Lendasse. Machine learning techniques based onrandom projections. In Michel Verleysen, editor, ESANN2010: 18th European Symposium onArtificial Neural Networks, Computational Intelligence and Machine Learning, pages 295–302. d-side Publications, Bruges, Belgium, April 28–30 2010.

[66] Elina Parviainen, Jaakko Riihimäki, Yoan Miche and Amaury Lendasse. Interpreting extremelearning machine as an approximation to an infinite neural network. In KDIR 2010: Proceedings ofthe International Conference on Knowledge Discovery and Information Retrieval. Valencia, Spain,October 2010.

[67] Federico Montesino Pouzols and Amaury Lendasse. Effect of different detrending approaches oncomputational intelligence models of time series. In International Joint Conference on NeuralNetworks (IJCNN), pages 1729–1736. Barcelona, Spain, July 2010.

[68] Federico Montesino Pouzols and Amaury Lendasse. Evolving fuzzy optimally pruned extreme learn-ing machine: A comparative analysis. In IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), pages 1339–1346. Barcelona, Spain, July 2010.

[69] Antti Sorjamaa, Amaury Lendasse and Eric Séverin. Combination of SOMs for fast missing valueimputation. In Proceedings of MASHS 2010, Modèles et Apprentissage en Sciences Humaines etSociale, Lille (France). Models and learnings in Human and social Sciences, June 2010.

[70] Mark van Heeswijk, Yoan Miche, Erkki Oja and Amaury Lendasse. Solving large regression prob-lems using an ensemble of GPU-accelerated ELMs. In Michel Verleysen, editor, ESANN2010:18th European Symposium on Artificial Neural Networks, Computational Intelligence and MachineLearning, pages 309–314. d-side Publications, Bruges, Belgium, April 28–30 2010.

[71] Francesco Corona, Elia Liitiäinen, Amaury Lendasse, Roberto Baratti and Lorenzo Sassu. De-launay tessellation and topological regression: An application to estimating product properties. InEvaristo Biscaia Rita de Brito Alves, Claudio Oller do Nascimento, editor, Computer Aided Chem-ical Engineering: Proceedings of PSE 2009 International Symposium on Process Systems Engineer-ing, Salvador Bahia (Brazil), Computer Aided Chemical Engineering, volume 27, pages 1179–1184.Elsevier, August 16-20 2009. doi:10.1016/S1570-7946(09)70587-5.

[72] Alberto Guillén, Antti Sorjamaa, Yoan Miche, Amaury Lendasse and Ignacio Rojas. Efficientparallel feature selection for steganography problems. In J. Cabestany, F. Sandoval A. Prieto andJ.M. Corchado, editors, LNCS - Bio-Inspired Systems: Computational and Ambient Intelligence ŰIWANN 2009, Part I, Lecture Notes in Computer Science, volume 5517/2009, page 1224 Ű 1231.IWANN, Springer Berlin / Heidelberg, June 2009. doi:10.1007/978-3-642-02478-8_153.

5

[73] Alberto Guillén, Antti Sorjamaa, Gines Rubio, Amaury Lendasse and Ignacio Rojas. Mutual infor-mation based initialization of forward-backward search for feature selection in regression problems.In C. Alippi, M.M. Polycarpou, C. Panayiotou and G. Ellinas, editors, LNCS - Artificial NeuralNetworks - ICANN 2009 Ű Part I, Lecture Notes in Computer Science, volume 5768, pages 1–9.ICANN, Springer Berlin / Heidelberg, September 2009. doi:10.1007/978-3-642-04274-4.

[74] Elia Liitiäinen, Amaury Lendasse and Francesco Corona. On the statistical estimation of rényientropies. In Proceedings of IEEE/MLSP 2009 International Workshop on Machine Learning forSignal Processing, Grenoble (France). September 2-4 2009. doi:10.1109/MLSP.2009.5306242.

[75] Fernando Mateo, Dusan Sovilj, Rafael Gadea and Amaury Lendasse. RCGA-s/RCGA-SP methodsto minimize the delta test for regression tasks. In IWANN 2009, Lecture Notes in Computer Science,volume 5517, pages 359–366. Springer, Salamanca, Spain, June 10-12 2009. doi:10.1007/978-3-642-02478-8.

[76] Paul Merlin, Antti Sorjamaa, Bertrand Maillet and Amaury Lendasse. X-SOM and l-SOM: anested approach for missing value imputation. In Michel Verleysen, editor, ESANN2009 proceedings,European Symposium on Artificial Neural Networks - Advances in Computational Intelligence andLearning, ESANN Proceedings, pages 83–88. ESANN, d-side publications, Brugge, Belgium, April2009.

[77] Yoan Miche and Amaury Lendasse. A faster model selection criterion for OP-ELM and OP-KNN: Hannan-quinn criterion. In Michel Verleysen, editor, ESANN’09: European Symposium onArtificial Neural Networks, pages 177–182. d-side publications, April 22-24 2009.

[78] Antti Sorjamaa, Francesco Corona, Yoan Miche, Paul Merlin, Bertrand Maillet and AmauryLendasse. Linear combination of SOMs for data imputation: Application to financial problems.In Proceedings of MASHS 2009, Modèles et Apprentissage en Sciences Humaines et Sociale, Lille(France). MASHS, June 8-9 2009.

[79] Antti Sorjamaa, Francesco Corona, Yoan Miche, Paul Merlin, Bertrand Maillet, Eric Séverin andAmaury Lendasse. Sparse linear combination of SOMs for data imputation: Application to fi-nancial database. In Risto Principe, J.C.; Miikkulainen, editor, Lecture Notes in Computer Sci-ence: Advances in Self-Organizing Maps - Proceedings of WSOM 2009 International Workshopon Self-Organizing Maps, Saint Augustine (Florida), Lecture Notes in Computer Science, volume5629/2009, pages 290–297. Springer Berlin / Heidelberg, June 8-10 2009. doi:10.1007/978-3-642-02397-2_33.

[80] Antti Sorjamaa, Paul Merlin, Bertrand Maillet and Amaury Lendasse. A non-linear approachfor completing missing values in temporal databases. European Journal of Economic and SocialSystems, volume 22(1):pages 99–117, November 2009. doi:10.3166/EJESS.22.99-117.

[81] Souhaib Ben Taieb, Gianluca Bontempi, Antti Sorjamaa and Amaury Lendasse. Long-term pre-diction of time series by combining direct and MIMO strategies. In International Joint Conferenceon Neural Networks. Atlanta, Georgia, USA, June 2009.

[82] Mark van Heeswijk, Yoan Miche, Tiina Lindh-Knuutila, Peter A.J. Hilbers, Timo Honkela, ErkkiOja and Amaury Lendasse. Adaptive ensemble models of extreme learning machines for time seriesprediction. In Cesare Alippi, Marios M. Polycarpou, Christos G. Panayiotou and Georgios Ellinas,editors, ICANN (2), Lecture Notes in Computer Science, volume 5769, pages 305–314. Springer,2009. doi:10.1007/978-3-642-04277-5_31.

[83] Qi Yu, Amaury Lendasse and Eric Séverin. Ensemble KNNs for bankruptcy prediction. In CEF 09,15th International Conference: Computing in Economics and Finance, Sydney. June 15-17 2009.

[84] Emil Eirola, Elia Liitiäinen, Amaury Lendasse, Francesco Corona and Michel Verleysen. Using thedelta test for variable selection. In M. Verleysen, editor, Proceedings of ESANN 2008, EuropeanSymposium on Artificial Neural Networks, Bruges (Belgium), pages 25–30. d-side publ. (Evere,Belgium), April 23-25 2008.

6

[85] Alberto Guillén, Luis Herrera, Gines Rubio, Amaury Lendasse, Hector Pomares and Ignacio Rojas.Instance or prototype selection for function approximation using mutual information. In AmauryLendasse, editor, ESTSP’08 Proceedings, pages 67–75. September 2008.

[86] Alberto Guillén, Dusan Sovilj, Fernando Mateo, Ignacio Rojas and Amaury Lendasse. New method-ologies based on delta test for variable selection in regression problems. In Workshop on ParallelArchitectures and Bioinspired Algorithms. Toronto, Canada, October 25-29 2008.

[87] Amaury Lendasse and Francesco Corona. Linear projection based on noise variance estimation:Application to spectral data. In M. Verleysen, editor, Proceedings of ESANN 2008, EuropeanSymposium on Artificial Neural Networks, Bruges (Belgium), pages 457–462. d-side publ. (Evere,Belgium), April 23-25 2008.

[88] Fernando Mateo and Amaury Lendasse. A variable selection approach based on the delta test forextreme learning machine models. In M. Verleysen, editor, Proceedings of the European Symposiumon Time Series Prediction, pages 57–66. d-side publ. (Evere, Belgium), September 2008.

[89] Yoan Miche, Patrick Bas, Christian Jutten, Olli Simula and Amaury Lendasse. A methodologyfor building regression models using extreme learning machine: OP-ELM. In M. Verleysen, editor,ESANN 2008, European Symposium on Artificial Neural Networks, Bruges, Belgium, pages 247–252. d-side publ. (Evere, Belgium), April 23-25 2008.

[90] Yoan Miche, Antti Sorjamaa and Amaury Lendasse. OP-ELM: Theory, experiments and a toolbox.In Roman Neruda Vera Kurková and Jan Koutník, editors, LNCS - Artificial Neural Networks- ICANN 2008 - Part I, Lecture Notes in Computer Science, volume 5163/2008, pages 145–154.Springer Berlin / Heidelberg, September 2008. doi:10.1007/978-3-540-87536-9_16.

[91] Federico Montesino Pouzols, Amaury Lendasse and Angel Barriga Barros. Fuzzy inferencebased autoregressors for time series prediction using nonparametric residual variance estima-tion. In 17th IEEE International Conference on Fuzzy Systems (FUZZ-IEEE’2008), IEEE WorldCongress on Computational Intelligence, pages 613–618. Hong Kong, China, June 2008. doi:10.1109/FUZZY.2008.4630432.

[92] Federico Montesino Pouzols, Amaury Lendasse and Angel Barriga Barros. xftsp: a tool for timeseries prediction by means of fuzzy inference systems. In 4th IEEE International Conferenceon Intelligent Systems (IS08), volume 1, pages 2–2–2–7. Varna, Bulgaria, September 2008. doi:10.1109/IS.2008.4670398.

[93] Risto Ritala, Esa Alhoniemi, Tuomo Kauranne, Kimmo Konkarikoski, Amaury Lendasse and MikiSirola. Nonlinear temporal and spatial forecasting: modelling and uncertainty analysis (notes) ŰMASIT20. In MASI Programme 2005Ű2009,Yearbook 2008, pages 163–175. 2008.

[94] Olli Simula, Francesco Corona, Amaury Lendasse, Marja-Liisa Riekkola, Kari Hartonen, PenttiMinkkinen, Satu-Pia Reinikainen, Jarno Kohonen, Ilppo Vuorinen, Jari Hänninen and Jukka Silén.Developing chemometrics with the tools of information sciences (CHESS) – MASIT23. In MASIProgramme 2005-2009, Yearbook 2008, pages 189–222. Libris Oy, May 2008.

[95] Antti Sorjamaa, Yoan Miche, Robert Weiss and Amaury Lendasse. Long-term prediction of timeseries using NNE-based projection and OP-ELM. In IEEE World Conference on ComputationalIntelligence, pages 2675–2681. Research Publishing Services, Chennai, India, Hong Kong, June2008. doi:10.1109/IJCNN.2008.4634173.

[96] Qi Yu, Antti Sorjamaa, Yoan Miche, Amaury Lendasse, Alberto Guillén, Eric Séverin and FernandoMateo. Optimal pruned k-nearest neighbors: OP-KNN - application to financial modeling. In FatosXhafa, Francisco Herrera, Ajith Abraham, Mario Köppen and Jose Manuel Bénitez, editors, HybridIntelligent Systems, 2008. Eighth International Conference on, pages 764–769. Barcelona, Spain,September 2008. doi:10.1109/HIS.2008.134.

[97] Qi Yu, Antti Sorjamaa, Yoan Miche, Eric Séverin and Amaury Lendasse. OP-KNN for financialregression problems. In Mashs 08, Computational Methods for Modelling and learning in Socialand Human Sciences, Creteil (France). June 5-6 2008.

7

[98] Francesco Corona and Amaury Lendasse. Variable scaling for time series prediction. In Proceedingsof ESTSP 2007, European Symposium on Time Series Prediction, Espoo (Finland), pages 69–76.February 7-9 2007.

[99] Francesco Corona, Elia Liitiäinen and Amaury Lendasse. Using functional representations in spec-trophotoscopic variables selection and regression. In Proceedings of SSC10 Scandinavian Symposiumon Chemometrics, Lappeenranta (Finland), page 29. June 11-15 2007.

[100] Francesco Corona, Elia Liitiäinen, Amaury Lendasse and Roberto Baratti. Measures of topologicalrelevance based on the self-organizing map: Applications to process monitoring from spectroscopicmeasurements. In Proceedings of EANN 2007, International Conference on Engineering Applica-tions of Neural Networks, Thessaloniki (Greece), pages 24–33. August 29-31 2007.

[101] Tuomas Kärnä, Francesco Corona and Amaury Lendasse. Compressing spectral data using opti-mised gaussian basis. In Proceedings of Chimiométrie 2007, Lyon (France), pages 177–180. Novem-ber 29-30 2007.

[102] Tuomas Kärnä and Amaury Lendasse. Comparison of FDA based time series prediction methods.In ESTSP 2007, European Symposium on Time Series Prediction, Espoo (Finland), pages 77–86.February 7-9 2007.

[103] Tuomas Kärnä and Amaury Lendasse. Gaussian fitting based FDA for chemometrics. In Fran-cisco Sandoval et al., editor, IWANN’07, International Work-Conference on Artificial Neural Net-works, San Sebastian, Spain, Lecture Notes in Computer Science, volume 4507, pages 186–193.Springer Berlin / Heidelberg, June 2007. doi:10.1007/978-3-540-73007-1_23.

[104] Tuomas Kärnä and Amaury Lendasse. Optimal gaussian basis functions for chemometrics. InSSC10, 10th Scandinavian Symposium on Chemometrics, Lappeenranta (Finland), page 79. June11-15 2007.

[105] Amaury Lendasse and Francesco Corona. Optimal linear projection based on noise variance es-timation. In Proceedings of Chimiométrie 2007, Lyon (France), pages 165–168. November 29-302007.

[106] Amaury Lendasse and Francesco Corona. Optimal linear projection based on noise variance esti-mation: Application to spectrometric modeling. In Proceedings of SSC10 Scandinavian Symposiumon Chemometrics, Lappeenranta (Finland), page 26. June 11-15 2007.

[107] Amaury Lendasse, Francesco Corona, Satu-Pia Reinikainen and Pentti Minkkinen. Functionalvariable selection using noise variance estimation. In Proceedings of Chimiométrie 2007, Lyon(France), pages 39–42. November 29-30 2007.

[108] Elia Liitiäinen, Francesco Corona and Amaury Lendasse. Nearest neighbor distributions and noisevariance estimation. In M. Verleysen, editor, Proceedings of ESANN 2007, European Symposiumon Artificial Neural Networks, Bruges (Belgium), pages 67–72. d-side publ. (Evere, Belgium), April25-27 2007.

[109] Elia Liitiäinen, Francesco Corona and Amaury Lendasse. Non-parametric residual variance esti-mation in supervised learning. In Francisco Sandoval et al., editor, Lecture Notes in ComputerScience: Computational and Ambient Intelligence - Proceedings of IWANN 2007 InternationalWork-Conference on Artificial Neural Networks, San Sebastian (Spain), Lecture Notes in ComputerScience, volume 4507/2007, pages 63–71. Springer-Verlag, June 20-22 2007. doi:10.1007/978-3-540-73007-1_9.

[110] Elia Liitiäinen and Amaury Lendasse. Variable scaling for time series prediction: Application to theESTSP’07 and the NN3 forecasting competitions. In IJCNN 2007, International Joint Conferenceon Neural Networks, Orlando, Florida, USA, pages 2812 – 2816. Documation LLC, Eau Claire,Wisconsin, USA, August 2007. doi:10.1109/IJCNN.2007.4371405.

8

[111] Yoan Miche, Patrick Bas, Amaury Lendasse, Christian Jutten and Olli Simula. Advantages ofusing feature selection techniques on steganalysis schemes. In Francisco Sandoval et al., editor,IWANN’07: International Work-Conference on Artificial Neural Networks, San Sebastian, Spain,Lecture Notes in Computer Science, volume 4507/2007, pages 606–613. Springer Berlin / Heidel-berg, June 20-22 2007. doi:10.1007/978-3-540-73007-1_73.

[112] Yoan Miche, Patrick Bas, Amaury Lendasse, Christian Jutten and Olli Simula. Extracting relevantfeatures of steganographic schemes by feature selection techniques. In Wacha’07: Third WavillaChallenge. Saint-Malo, France, June 14 2007.

[113] Yoan Miche, Patrick Bas, Amaury Lendasse, Olli Simula and Christian Jutten. Avantages de lasélection de caractéristiques pour la stéganalyse. In GRETSI 2007, Groupe de Recherche et d’Etudesdu Traitement du Signal et des Images, Troyes, France. Troyes, France, September 11-13 2007.

[114] Nima Reyhani and Amaury Lendasse. An empirical dependence measures based on residual varianceestimation. In ISSPA 2007, International Symposium on Signal Processing and its Applicationsin conjunction with the International Conference on Information Sciences, Signal Processing andits Applications, Sharjah, United Arab Emirates (U.A.E.), pages 1–4. February 12-15 2007. doi:10.1109/ISSPA.2007.4555501.

[115] Olli Simula, Amaury Lendasse, Francesco Corona, Satu-Pia Reinikainen, Marja-Liisa Riekkola,Kari Hartonen, Ilppo Vuorinen and Jukka Silén. Developing chemometrics with the tools of infor-mation sciences (CHESS) Ű MASIT23. In MASI Programme 2005-2009, Yearbook 2007, pages201–221. Libris Oy, March 2007.

[116] Antti Sorjamaa and Amaury Lendasse. Time series prediction as a problem of missing values. InESTSP 2007, European Symposium on Time Series Prediction, Espoo (Finland), pages 165–174.February 7-9 2007.

[117] Antti Sorjamaa, Elia Liitiäinen and Amaury Lendasse. Time series prediction as a problem ofmissing values: Application to ESTSP2007 and NN3 competition benchmarks. In IJCNN, Inter-national Joint Conference on Neural Networks, pages 1770–1775. Documation LLC, Eau Claire,Wisconsin, USA, Orlando, Florida, USA, August 12-17 2007. doi:10.1109/IJCNN.2007.4371429.

[118] Antti Sorjamaa, Paul Merlin, Bertrand Maillet and Amaury Lendasse. A nonlinear approach forthe determination of missing values in temporal databases. In MASHS, Computational Methods forModelling and learning in Social and Human Sciences, Brest (France). May 10-11 2007.

[119] Antti Sorjamaa, Paul Merlin, Bertrand Maillet and Amaury Lendasse. SOM+EOF for findingmissing values. In M. Verleysen, editor, ESANN 2007, European Symposium on Artificial NeuralNetworks, Bruges (Belgium), pages 115–120. d-side publ. (Evere, Belgium), April 25-27 2007.

[120] J. Vandewalle, J. Suykens, B. De Moor and Amaury Lendasse. State-of-the-art and evolution inpublic data sets and competitions for system identification, time series prediction and pattern recog-nition. In 32nd International Conference on Acoustics, Speech, and Signal Processing (ICASSP),Hawaii Convention Center in Honolulu (USA), volume 4, pages 1269–1272. April 15-20 2007.doi:10.1109/ICASSP.2007.367308.

[121] Qi Yu, Eric Séverin and Amaury Lendasse. A global methodology for variable selection: Applicationto financial modeling. In Mashs 2007, Computational Methods for Modelling and learning in Socialand Human Sciences, Brest (France). May 10-11 2007.

[122] Qi Yu, Eric Séverin and Amaury Lendasse. Variable selection for financial modeling. In CEF2007, 13th International Conference on Computing in Economics and Finance MontrÃľal, Quebec,Canada. June 14 -16 2007.

[123] Amaury Lendasse, Francesco Corona, Jin Hao, Nima Reyhani and Michel Verleysen. Determina-tion of the mahalanobis matrix using non-parametric noise estimations. In M. Verleysen, editor,Proceedings of ESANN 2006, European Symposium on Artificial Neural Networks, Bruges (Lille),pages 227–232. Bruges, Belgium, April 26-28 2006.

9

[124] Elia Liitiäinen and Amaury Lendasse. Long-term prediction of time series using state-space models.In S. Kollias et al., editor, ICANN’06, International Conference on Artificial Neural Networks, 16thInternational Conference, Athens (Greece), Lecture Notes in Computer Science, volume 4132/2006,pages 181–190. Springer-Verlag, September 10-14 2006. doi:10.1007/11840930_19.

[125] Yoan Miche, Benoit Roue, Patrick Bas and Amaury Lendasse. A feature selection methodology forsteganalysis. In B. Gunsel, A. K. Jain, A. M. Tekalp and B. Sankur, editors, MRCS06, InternationalWorkshop on Multimedia Content Representation, Classification and Security, Istanbul (Turkey),Lecture Notes in Computer Science, volume 4105, pages 49–56. Springer-Verlag, September 11-132006. doi:10.1007/11848035_9.

[126] Antti Sorjamaa and Amaury Lendasse. Time series prediction using dirrec strategy. In M. Ver-leysen, editor, ESANN06, European Symposium on Artificial Neural Networks, pages 143–148.European Symposium on Artificial Neural Networks, Bruges, Belgium, April 26-28 2006.

[127] Jarkko Tikka, Amaury Lendasse and Jaakko Hollmén. Analysis of fast input selection: Applica-tion in time series prediction. In S. Kollias et al., editor, ICANN06, International Conferenceon Artificial Neural Networks, 16th International Conference, Athens (Greece), Lecture Notes inComputer Science, volume 4132/2006, pages 161–170. Springer-Verlag, September 10-14 2006. doi:10.1007/11840930_17.

[128] Francesco Corona and Amaury Lendasse. Input selection and function approximation using theself-organizing map: An application to spectrometric modeling. In Proceedings of WSOM 2005International Workshop on Self-Organizing Maps, Paris (France), pages 653–660. September 5-82005.

[129] Yongnan Ji, Jin Hao, Nima Reyhani and Amaury Lendasse. Direct and recursive prediction of timeseries using mutual information selection. In J. Cabestany et al., editor, Computational Intelligenceand Bioinspired Systems: 8th International Workshop on Artificial Neural Networks, IWANN’05,Vilanova i la Geltra, Barcelona, Spain, Lecture Notes in Computer Science, volume 3512, pages1010–1017. Springer-Verlag GmbH, June 8-10 2005. doi:10.1007/11494669_124.

[130] Amaury Lendasse, Damien François, Vincent Wertz and Michel Verleysen. Nonparametric noiseestimation to build nonlinear model in chemometry. In Chimiométrie 2005, Villeneuve d’Ascq(France), pages 143–146. November 30 - December 1 2005.

[131] Amaury Lendasse, Yongnan Ji, Nima Reyhani and Michel Verleysen. LS-SVM hyperparameterselection with a nonparametric noise estimator. In W. Duch et al., editor, ICANN05, InternationalConference on Artificial Neural Networks, Artificial Neural Networks: Formal Models and TheirApplications, Lecture Notes in Computer Science, volume 3697, pages 625–630. September 11-152005.

[132] Nima Reyhani, Jin Hao, Yongnan Ji and Amaury Lendasse. Mutual information and gamma testfor input selection. In M. Verleysen, editor, ESANN 2005, European Symposium on ArtificialNeural Networks, Bruges (Belgium), pages 503–508. d-side publ. (Evere, Belgium), April 27-292005.

[133] Antti Sorjamaa, Jin Hao and Amaury Lendasse. Mutual information and k-nearest neighborsapproximator for time series predictions. In W. Duch at al., editor, LNCS - Artificial NeuralNetworks: Formal Models and Their Applications - ICANN 2005, Lecture Notes in ComputerScience, volume 3697/2005, pages 553–558. Springer Berlin / Heidelberg, September 11-15 2005.doi:10.1007/11550907_87.

[134] Antti Sorjamaa, Amaury Lendasse and Michel Verleysen. Pruned lazy learning models for timeseries prediction. In M. Verleysen, editor, ESANN05, European Symposium on Artificial NeuralNetworks, pages 509–514. d-side publ. (Evere, Belgium), April 27-29 2005.

[135] Antti Sorjamaa, Nima Reyhani and Amaury Lendasse. Input and structure selection for k-NN ap-proximator. In Francisco Sandoval Joan Cabestany, Alberto Prieto, editor, LNCS - ComputationalIntelligence and Bioinspired Systems - IWANN 2005, Lecture Notes in Computer Science, volume3512/2005, pages 985–992. Springer Berlin / Heidelberg, June 2005. doi:10.1007/11494669_121.

10

[136] Jarkko Tikka, Jaakko Hollmén and Amaury Lendasse. Input selection for long-term predictionof time series. In Francisco Sandoval Joan Cabestany, Alberto Prieto, editor, ComputationalIntelligence and Bioinspired Systems: 8th International Workshop on Artificial Neural Networks,IWANN 2005, Vilanova i la Geltra, Barcelona, Spain, Lecture Notes in Computer Science, volume3512, pages 1002–1009. Springer-Verlag GmbH, June 2005. ISBN 3-540-26208-3. ISSN 0302-9743.

[137] Igor Beliaev, Robert Kozma and Amaury Lendasse. Robust time series prediction using KIIImodel. In IDS04 Symposium, FedEx Institute of Technology (FIT), University of Memphis, TN,USA, pages April 24–26. Published 2004.

[138] Amaury Lendasse, Damien François, Fabrice Rossi, Vincent Wertz and Michel Verleysen. Sélectionde variables spectrales par information mutuelle multivariée pour la construction de modèles non-linéaires. In Chimiométrie 2004, Paris (France), pages 44–47. November 30 - December 1 2004.

[139] Amaury Lendasse, Erkki Oja, Olli Simula and Michel Verleysen. Time series prediction competi-tion: The CATS benchmark. In IJCNN 2004, International Joint Conference on Neural Networks,volume 2, pages 1615–1620. Budapest, Hungary, July, 25-29 2004.

[140] Amaury Lendasse, Geoffroy Simon, Robert Kozma, Vincent Wertz and Michel Verleysen. Fastbootstrap for least-square support vector machines. In M. Verleysen, editor, ESANN 2004, EuropeanSymposium on Artificial Neural Networks, Bruges (Belgium), pages 525–530. d-side publ. (Evere,Belgium), April 28-30 2004.

[141] Amaury Lendasse, Vincent Wertz, Geoffroy Simon and Michel Verleysen. Fast bootstrap appliedto LS-SVM for long term prediction of time series. In Neural Networks, 2004. Proceedings. 2004IEEE International Joint Conference on, volume 1, pages 705–710. IEEE, July 2004.

[142] Antti Sorjamaa, Amaury Lendasse, Damien François and Michel Verleysen. Business plans clas-sification with locally pruned lazy learning models. In ACSEG 2004, Connectionist Approaches inEconomics and Management Sciences, Lille (France), pages 112–119. November 18-19 2004.

[143] Simon Dablemont, Geoffroy Simon, Amaury Lendasse, Alain Ruttiens, François Blayo and MichelVerleysen. Time series forecasting with SOM and local non-linear models - application to theDAX30 index prediction. In Proceedings of the Workshop on Self-organizing Maps, pages 340–345.Hibikino, Japan, September 11-14 2003.

[144] Simon Dablemont, Geoffroy Simon, Amaury Lendasse, Alain Ruttiens and Michel Verleysen. Fi-nancial time series forecasting by double SOM maps and local RBF models forecasting the DAX30index. In ACSEG 2003, Rencontre Internationale sur les Approches Connexionnistes en SciencesEconomiques et de Gestion, Nantes (France), pages 153–164. November 20-21 2003.

[145] Damien François, Benoit Gailly, Amaury Lendasse, Vincent Wertz and Michel Verleysen. Shouldseed investors read business plans? In 22th Benelux Meeting on Systems and Control, Lommel,Belgium. March 19-21 2003.

[146] Damien François, Amaury Lendasse, Benoit Gailly, Vincent Wertz and Michel Verleysen. Arebusiness plans usefull for investors ? In ACSEG 2003, Connectionist Approaches in Economicsand Management Sciences, Nantes (France), pages 239–249. November 20-21 2003.

[147] Amaury Lendasse, Damien François, Vincent Wertz and Michel Verleysen. Nonlinear time seriesprediction by weighted vector quantization. In P.M.A. Sloot et al., editor, Computational ScienceŮ ICCS 2003, Lecture Notes in Computer Science, volume 2657–1, pages 417–426. Springer Berlin/ Heidelberg, January 2003. doi:10.1007/3-540-44860-8.

[148] Amaury Lendasse, John A. Lee, Eric de Bodt, Vincent Wertz and Michel Verleysen. Approximationby Radial-Basis Function networks - Application to option pricing, Advances in ComputationalManagement Science, C. Lesage and M. Cottrell editors, volume 6, chapter 10 in ConnectionistApproaches in Economics and Management Sciences, pages 203–214. Kluwer academic, 2003.

11

[149] Amaury Lendasse, Geoffroy Simon, Vincent Wertz and Michel Verleysen. Fast bootstrap for modelstructure selection. In 22th Benelux Meeting on Systems and Control, Lommel, Belgium, page 81.March 19-21 2003.

[150] Amaury Lendasse, Vincent Wertz and Michel Verleysen. Model selection with cross-validations andbootstraps - application to time series prediction with RBFN models. In O. Kaynak, E. Alpaydin,E. Oja and L. Xu, editors, ICANN 2003, Joint International Conference on Artificial NeuralNetworks, Istanbul (Turkey), Lecture Notes in Computer Science, volume 2714, pages 573–580.Springer-Verlag, June 26-29 2003. doi:10.1007/3-540-44989-2_68.

[151] Geoffroy Simon, Amaury Lendasse, Marie Cottrell, Jean-Claude Fort and Michel Verleysen. DoubleSOM for long-term time series prediction. In WSOM 2003, Workshop on Self-Organizing Maps,pages 35–40. Hibikino, Japan, September 11-14 2003.

[152] Geoffroy Simon, Amaury Lendasse, Marie Cottrell and Michel Verleysen. Long-term time seriesforecasting using self-organizing maps: the double vector quantization method. In ANNPR 2003,Artificial Neural Networks in Pattern Recognition, Florence (Italy), pages 8–14. September 12-132003.

[153] Geoffroy Simon, Amaury Lendasse and Michel Verleysen. Bootstrap for model selection: Linearapproximation of the optimism. In J.R. Alvarez J. Mira, editor, IWANN 2003, International Work-Conference on Artificial and Natural Neural Networks, Mao, Menorca (Spain), Lecture Notes inComputer Science, volume 2686–1, pages 182–189. Springer-Verlag, June 3-6 2003. doi:10.1007/3-540-44868-3_24.

[154] Geoffroy Simon, Amaury Lendasse, Vincent Wertz and Michel Verleysen. Fast approximation ofthe bootstrap for model selection. In M. Verleysen, editor, ESANN 2003, European Symposium onArtificial Neural Networks, Bruges (Belgium), pages 99–106. d-side publ. (Evere, Belgium), April23-25 2003.

[155] Michel Verleysen and Amaury Lendasse. Le test des méthodes neuronales Ű ou comment utiliserles techniques de rééchantillonnage pour ne pas se tromper de résultat. In ACSEG 2003 proceedings- Connectionist Approaches in Economics and Management Sciences, Nantes (France), pages 515–534. November 20-21 2003.

[156] Nabil Benoudjit, Cédric Archambeau, Amaury Lendasse, John A. Lee and Michel Verleysen. Widthoptimization of the gaussian kernels in radial basis function networks. In M. Verleysen, editor,ESANN 2002, European Symposium on Artificial Neural Networks, Bruges (Belgium), pages 425–432. d-side publ. (Evere, Belgium), April 2002.

[157] Pierre Cardon, Amaury Lendasse, Vincent Wertz, Eric de Bodt and Michel Verleysen. Classifi-cation of investment funds by self-organizing maps. In ACSEG 2002, Connectionist Approachesin Economics and Management Sciences, Boulogne-sur-Mer (France), pages 201–212. November21-22 2002.

[158] Amaury Lendasse, Marie Cottrell, Vincent Wertz and Michel Verleysen. Prediction of electric loadusing kohonen maps - application to the polish electricity consumption. In ACC 2002, AmericanControl Conference, Anchorage, Alaska (USA), pages 3684–3689. June 2002.

[159] Amaury Lendasse and Michel Verleysen. Curvilinear distance analysis versus isomap. In M. Verley-sen, editor, ESANN 2002, European Symposium on Artificial Neural Networks, Bruges (Belgium),pages 185–192. d-side publ. (Evere, Belgium), April 2002.

[160] Cédric Archambeau, Amaury Lendasse, Charles Trullemans, Claude Veraart, Jean Delbeke andMichel Verleysen. Phosphene evaluation in a visual prosthesis with artificial neural networks. InAdaptive Systems and Hybrid Computational Intelligence in Medicine, special session proceedingsof EUNITE 2001, Tenerife (Spain), pages 116–122. December 13-14 2001.

12

[161] Cédric Archambeau, Amaury Lendasse, Charles Trullemans, Claude Veraart, Jean Delbeke andMichel Verleysen. Phosphene evaluation in a visual prosthesis with artificial neural networks.In EUNITE 2001, European Symposium on Intelligent Technologies, Hybrid Systems and theirimplementation on Smart Adaptive Systems, Tenerife (Spain), pages 509–515. December 13-142001.

[162] Amaury Lendasse, John A. Lee, Eric de Bodt, Vincent Wertz and Michel Verleysen. Approximationusing radial basis functions networks - application to pricing derivative securities. In ACSEG 2001,Connectionist Approaches in Economics and Management Sciences, Rennes (France), pages 275–283. November 22-23 2001.

[163] Amaury Lendasse, John A. Lee, Eric de Bodt, Vincent Wertz and Michel Verleysen. Input datareduction for the prediction of financial time series. In M. Verleysen, editor, ESANN 2001, Eu-ropean Symposium on Artificial Neural Networks, Bruges (Belgique), pages 237–244. d-side publ.(Evere, Belgium), April 2001.

[164] Amaury Lendasse, Vincent Wertz and Michel Verleysen. Forecasting electricity demand usingkohonen maps. In 20th Benelux meeting on Systems and Control, Houffalize (Belgium), page 118.March 2001.

[165] G. Gomez and Amaury Lendasse. Statistical fault isolation with PCA. In IFAC, Safeprocess’. 2000.

[166] John A. Lee, Amaury Lendasse, N. Donckers and Michel Verleysen. A robust non-linear projectionmethod. In M. Verleysen, editor, ESANN’2000, European Symposium on Artificial Neural Networks,Bruges (Belgique), pages 13–20. April 2000.

[167] Amaury Lendasse, John A. Lee, Eric de Bodt, Vincent Wertz and Michel Verleysen. Réduction dela dimension d’un ensemble d’indicateurs techniques en vue de la prédiction de séries temporellesfinancières - application à l’indice de marché BEL 20. In ACSEG 2000, 7emes rencontres inter-nationales. December 2000.

[168] Amaury Lendasse, John A. Lee, Vincent Wertz and Michel Verleysen. Time series forecastingusing CCA and kohonen maps - application to electricity consumption. In M. Verleysen, editor,ESANN’2000, European Symposium on Artificial Neural Networks, Bruges (Belgique), pages 329–334. April 2000.

[169] N. Donckers, Amaury Lendasse, Vincent Wertz and Michel Verleysen. Extraction of intrinsic di-mension using CCA - application to blind sources separation. In ESANN’99, European Symposiumon Artificial Neural Networks, Bruges (Belgique), pages 339–344. April 1999.

[170] Amaury Lendasse. Comparison between NAR and NARMA models for time-series prediction:Choice of a non-linear regressor vector. In 18th Benelux Meeting on Systems and Control, Confer-ence Center "Hengelhoef", Houthalen, Belgium. March 3-5 1999.

[171] Michel Verleysen, Eric de Bodt and Amaury Lendasse. Forecasting financial time series throughintrinsic dimension estimation and non-linear data projection. In J. Sanchez-Andres J. Mira, editor,IWANN99, International Work-conference on Artificial and Natural Neural networks, Alicante(Spain). Published in Engineering Applications of Bio-Inspired Artificial Neural Networks, LectureNotes in Computer Science, volume 1607–2, pages 596–605. Springer-Verlag, June 1999. doi:10.1007/BFb0100527.

[172] M. L. Hadjili, Amaury Lendasse, Vincent Wertz and S. Yurkovich. Identification of fuzzy models fora glass furnace process. In 1998 IEEE International Conference on Control Applications,Trieste,Italy, pages 963–968. September 1-4 1998. doi:10.1109/CCA.1998.721601.

[173] Amaury Lendasse, Eric de Bodt and Michel Verleysen. Estimation de la dimension intrinsèqued’une série temporelle et prédiction par une méthode de projection. In ACSEG’98, AssociationConnectioniste en Sciences Economiques et de Gestion, Louvain-la-Neuve (Belgique), pages D37–D46. November 20 1998.

13

[174] Amaury Lendasse, Michel Verleysen, Eric de Bodt, Marie Cottrell and P. Gregoire. Forecastingtime-series by kohonen classification. In M. Verleysen, editor, ESANN’98, European Symposiumon Artificial Neural Networks, Bruges (Belgique), pages 221–226. April 1998.

Other Publications

[175] Yoan Miche, Patrick Bas and Amaury Lendasse. Using multiple re-embeddings for quantitativesteganalysis and image reliability estimation. Technical Report TKK-ICS-R34, Aalto UniversitySchool of Science and Technology, Aalto, Finland, June 2010.

[176] Antti Sorjamaa and Amaury Lendasse. Fast missing value imputation using ensemble of SOMs.Technical Report TKK-ICS-R33, Aalto University School of Science and Technology, June 2010.

[177] Amaury Lendasse, editor. ESTSP 2008: Proceedings. Multiprint Oy / Otamedia, 2008. ISBN:978-951-22-9544-9.

[178] Elia Liitiäinen, Francesco Corona and Amaury Lendasse. A boundary corrected expansion of themoments of nearest neighbor distributions. Technical Report TKK-ICS-R9, Helsinki University ofTechnology, October 18 2008.

[179] Amaury Lendasse, editor. ESTSP 2007: Proceedings. Multiprint Oy / Otamedia, 2007. ISBN:978-951-22-8601-0.

[180] Amaury Lendasse. Analyse et prédiction de séries temporelles par méthodes non linéaires: Appli-cation à des données industrielles et financières. Ph.D. thesis, Université catholique de Louvain,2003.

14